Hello all! I am Harsh Gupta. I am an Applied Scientist with the Amazon Search Science and Intelligence team. I defended my PhD thesis in the ECE department at the University of Illinois at Urbana-Champaign (UIUC) and graduated in December 2020! At UIUC, I had the privilege of working with Prof. R. Srikant and I was also affiliated with the Coordinated Science Laboratory (CSL).
I am primarily interested in the design and analysis of algorithms in reinforcement learning, multi-armed bandits (non-standard formulations) and optimization.
I am also interested in applying my theoretical research to practical problems arising in robotics/self-driving cars, virtual reality and networking.
Brief Overview of Research and Industry Experience:
I have done research in the analysis and design of different reinforcement learning algorithms (, , and ). I also have practical experience in applying techniques from reinforcement learning and model predictive control at Xsense.ai in order to improve the efficiency, comfort and safety of an autonomous vehicle merging on a highway.
I have done research in the analysis and design of algorithms for different instances of non-standard multi-armed bandit formulations (, , and a paper in the domain of virtual reality which is under submission). I have applied the designed algorithms to applications in wireless networks and virtual reality (see Publications).
I have done several internships applying machine learning techniques to practical problems arising in very different domains. These include the following: applying machine learning techniques to solve networking problems at Huawei, predicting crimes in Los Angeles to aid LAPD's initiatives for predictive policing at Celect (a start-up now acquired by Nike), improving the efficiency, safety and comfort of autonomous highway merging at Xsense.ai, and improving search product placement for Amazon's e-commerce website at Amazon Search Science and AI.
Additional details can be found under the different tabs on this website. Feel free to contact me with any questions, comments or suggestions!